KB Article #179285

Transfer CFT monitor becomes unresponsive after CFTH03E and CFTH07E

Problem

CFTH07E TFIL task Synchronization error


CFTH03E Error sending data on network



Resolution

This is usually related to an unpredictable TCP error or a too slow system / TCP that messed-up the Transfer CFT inter-task communication.


The issue can also be related to a temporary heavy loaded system with CFT not set as a Non-Swappable application.


To improve the performance Transfer CFT should be started as Started Task in a Non-Swappable TP class.

Moreover, to avoid a bottleneck with TCP-IP, CFT Started Task should be in the same performance group than TCP-IP Started Task.




Note: latest versions / SP includes a change to trap such a condition externally:

When a major TCP error was found, the Transfer CFT TCP task tried to stop but went into a loop.
A specific CTCP01F message is now sent for JESMSGLG, and a dump is processed (assert). Additionally, more messages can be found on TCPOUT.

Message text: CTCP01F Unpredictable TCP error, Forced Dump, monitor restart needed.